Who would have thought the Singularity could happen as a side effect of Siri's difficulty booking a table at a restaurant for the next free night on your schedule. It might, as Apple's next virtual assistant Viv may be able to "teach" itself to perform new tasks, eventually making itself smarter and smarter -- more like the AI depicted in "Her" and less like a simple voice-activated program.

Apple has been putting a small team on the task of creating an advanced artificial intelligence (AI) for its next generation of personal assistant software. And according to "Wired" -- which published an exclusive feature on the project and its competitors -- Apple's not alone.

The next generation of Apple's personal assistant software may be called "Viv," and it may be able to teach itself to learn new capabilities like a real AI. "Siri is chapter one of a much longer, bigger story," said Dag Kittlaus, who helped create Siri and is one of the team members working in a small startup called Viv Labs, to "Wired."

The problem with Siri as a personal assistant is that, behind the advanced language recognition and personable voice, it's still essentially an old-fashioned piece of coded software -- software that can do what it's programmed to do and nothing beyond. So, for example, Siri can book a restaurant reservation using the integrated OpenTable software, and it can check your calendar to see when you're free, but it can't book a table for the next evening you're free. To gain that capability, or an endless iteration of other useful tasks, like booking a flight or hotel room, Siri would have to be programmed to be able to do it.

Viv Labs -- whose website boasts the tagline "The Global Brain" -- is working on an assistant that would learn to do novel tasks by generating its own code as it needs to. The key to making it work is to give Viv access to as much data as possible, and then have "her" break down the natural language in novel queries or commands into distinct pieces that Viv can understand and look up, and then combine into an intelligent action that satisfies the request. Wired gives this example:

"'Give me a flight to Dallas with a seat that Shaq could fit in.' Viv will parse the sentence and then it will perform its best trick: automatically generating a quick, efficient program to link third-party sources of information together-say, Kayak, SeatGuru, and the NBA media guide-so it can identify available flights with lots of legroom. And it can do all of this in a fraction of a second."

Besides the concept and coding involved, a system like Viv would need a "boundless brain" that could quickly be trained to access specific databases and apps based on the real-world spoken language that refers to them. It's one of the three pillars, or principles, described by Viv's creators to "Wired": "Viv will be taught by the world, it will know more than it is taught, and it will learn something every day."

Those pillars basically describe the principle behind theoretical AI. The more users interact with the AI, the smarter it will become. Kittlaus is optimistic enough about the prospect that he sees a future scenario that could have been a scene from Spike Jonze's recent human/AI love story, "Her":

You're outside a dive bar at closing time. Viv "recognizes" your location and time -- and through the phone's accelerometers -- that you're holding the iPhone unsteadily. Without much more input than a phrase like "I'm drunk," Viv calls up your preferred car service to the bar and tells the driver "herself" the address to take you home to. (Note to Apple: try to get Scarlett Johansson to be the voice of Viv.)

If Viv sounds like a rare Apple moonshot project, it's not the only company working on a big leap in AI. Google recently bought a company called DeepMind and has worked with the likes of Ray Kurzweil at its headquarters, and Facebook is working on a so-called "deep-learning" project as well. And don't forget IBM, whose database-connected, natural language-processing brainbot Watson trounced the human world champion, Ken Jennings, at Jeopardy, causing him to personally welcome "our new computer overlords" when the last game's fate was sealed.

 

For more stories like this, follow us on Twitter!